AI Agent Access
Feature Introduction
Regardless of your programming experience, you can quickly set up various Q&A Agents based on AI models on the cloud console. Moreover, you can deploy the created Agents to Mini Program customer service, Service Accounts, Official Accounts, and WeChat Channels, allowing more users to chat with your AI Agents. We have released the AI Agent as an API service, allowing you to interact with the Agent via HTTP access.
This series of open APIs provides unified integration capabilities for AI Agents. Simply configure the token of the required Agent in the cloud development environment to quickly access the Agent. The conversation interface in this series supports responses in SSE format.
Access Guide
To call the following interfaces, you need to pass an AccessToken in the format Authorization: Bearer <token>. For how to obtain the token, refer to: Obtain token
HTTP Status Code
Depending on the returned results, status codes such as 2xx, 4xx, and 5xx may occur. The Agent indicates errors through the response body in 200 status codes, which should be handled accordingly.
Error Codes
The following lists the error codes specific to this series of APIs.
| Error Code | Meaning |
| AGENT_INVALID_PARAM | Invalid Agent call parameters |
| AGENT_GET_INFO_ERROR | Failed to query Agent information |
| AGENT_REQUEST_LLM_ERROR | Agent large model call exception |
| AGENT_REQUEST_KNOWLEDGE_ERROR | Agent Knowledge Base Query Exception |
| AGENT_SERVER_ERR | Agent System Exception |
Authentication
- HTTP: Bearer Auth
The token corresponding to the environment ID, obtained using login authentication (v2).
| Security Scheme Type: | http |
|---|---|
| HTTP Authorization Scheme: | bearer |